Choose the statement which best explains the construction of investment strategies based on autoregressive forecasting.

Select one:

a.

Using a universe of N assets and an estimation (i.e. look-back) period of the last K observations, we produce the AR(1) forecasts for the next period. Then, we can either build a directional investment strategy which opens long positions in those assets which have a positive sign estimate, and short position in those assets which have a negative sign estimate. Alternatively, we can build a cross-sectional investment strategy which opens long positions in the top Z assets with positive signs ranked according to their absolute forecasts, and short positions in the bottom Z assets with negative signs ranked according to their absolute forecasts.

b.

Using a universe of N assets and an estimation (i.e. look-back) period of the last K observations, we produce the AR(P) forecasts for the next period. Then, we can either build a directional investment strategy which opens long positions in those assets which have a positive sign estimate, and short position in those assets which have a negative sign estimate. Alternatively, we can build a cross-sectional investment strategy which opens long positions in the top Z assets with positive signs ranked according to their absolute forecasts, and short positions in the bottom Z assets with negative signs ranked according to their absolute forecasts.

c.

Using a universe of N assets and an estimation (i.e. look-back) period of the last K observations, we produce the AR(1) forecasts for the next period based on the AR(1) estimates. Then, we can either build a directional investment strategy which opens long positions in those assets which have a positive sign estimate, and short position in those assets which have a negative sign estimate. Alternatively, we can build a cross-sectional investment strategy which opens long positions in the top Z assets with negative signs ranked according to their forecasts, and short positions in the bottom Z assets with positive signs ranked according to their forecasts.

d.

Using a universe of N assets and an estimation (i.e. look-back) period of the last K observations, we produce the AR(P) forecasts for the next period based on the AR(P) estimates. Then, we can either build a directional investment strategy which opens long positions in those assets which have a positive sign estimate, and short position in those assets which have a negative sign estimate. Alternatively, we can build a cross-sectional investment strategy which opens long positions in the top Z assets with positive signs ranked according to their forecasts, and short positions in the bottom Z assets with negative signs ranked according to their forecasts.